Updating a site's engine

The person says: “The downtown site is still on the old engine; move it onto the current one.”

Sites move onto a new engine on their own, a few at a time, and a site whose move failed stays where it was until someone asks again. Moving one is platform work: it is asked for through work_request, runs in the background, and replaces what the site serves, so it needs an Admin’s yes.

The loop

1. Empress:search — query: "empress.capability_run": the platform work that moves a site, and what it takes.
2. Empress:work_request — operation: "empress.capability_run", sites (the site’s slug), args: { capability: "engine_roll" }; confirm: true once the person agreed. The answer is a task.
3. Empress:work_status — the task, until its state is done or failed.
4. Empress:site_status — the site, as it serves now.

Judgment

• The pages are checked before the move stands. A move whose pages do not render is put back on its own; a failed task names the site and why, and the site keeps serving what it served before.
• One site first. For several sites, move one, read its status, then name the rest in sites.
• Below Admin, the yes is someone else’s. The answer says the request waits in the console for an Admin; tell the person, and follow the task once it is approved.